边缘检测（8）Wallis微分算子

最新推荐文章于 2025-08-13 09:17:31 发布

原创最新推荐文章于 2025-08-13 09:17:31 发布 · 4.2k 阅读

12 ·

CC 4.0 BY-SA版权

图像处理同时被 2 个专栏收录

79 篇文章

订阅专栏

算法演化

57 篇文章

订阅专栏

本文介绍了一种结合Laplacian微分算子和对数算子的图像锐化技术——Wallis微分算子。该算子通过近似的对数运算，有效改善了暗部区域的锐化效果，特别适用于图像处理领域。

Wallis微分算子

模板为L=[ 0, -log, 0;

-log, 4log,-log;

0, -log, 0]

该模板是将Laplacian微分算子和对数算子结合的一种锐化算子。因为人眼对画面信号的处理过程中有一个近似的对数运算环节，解决了Laplacian微分算子对画面比较暗的部分锐化比较弱的缺陷。

%Wallis微分算子
clear all;
I=imread('lena.bmp');

%figure
%imshow(I);
I=im2double(I);

[height width R]=size(I);
for i=2:height-1
    for j=2:width-1
        II(i,j)=log10(I(i,j)+1)-0.25*(log10(I(i-1,j)+1)+log10(I(i+1,j)+1)+log10(I(i,j-1)+1)+log10(I(i,j+1)+1));
    end
end

min1=min(II);
min2=min(min1);
for i=2:height-1
   for j=2:width-1
      II(i,j)=46*II(i,j)-min2+0.4;
  end
end
%figure
%imshow(II,[]);


for i=1:height-1
    for j=1:width-1
        if (II(i,j)<-0.035)
           II(i,j)=0;
        else II(i,j)=1;
        end
    end
end
figure;
imshow(II,[]);

%Wallis微分算子
clear all;
I=imread('lena.bmp');

%figure
%imshow(I);
I=im2double(I);

[height width R]=size(I);
for i=2:height-1
    for j=2:width-1
        II(i,j)=log10(I(i,j)+1)-0.25*(log10(I(i-1,j)+1)+log10(I(i+1,j)+1)+log10(I(i,j-1)+1)+log10(I(i,j+1)+1));
    end
end

min1=min(II);
min2=min(min1);
for i=2:height-1
   for j=2:width-1
      II(i,j)=46*II(i,j)-min2+0.4;
  end
end
%figure
%imshow(II,[]);


for i=1:height-1
    for j=1:width-1
        if (II(i,j)<-0.035)
           II(i,j)=0;
        else II(i,j)=1;
        end
    end
end
figure;
imshow(II,[]);